openimageio: updated to 3.1.13.1
Release 3.1.13.1 and 3.1.13.0 (May 3, 2026) -- compared to 3.1.12.0
IBA: Add FLIP perceptual image difference metric as an experimental feature, including ImageBufAlgo::experimental::FLIP_diff() C++ API, Python ImageBufAlgo.FLIP_diff(), and oiiotool --flipdiff command (requires --experimental flag). Also introduces oiiotool --experimental flag to enable features not yet part of the stable API.
fmath.h: degrees() and radians() are now constexpr.
ImageSpec: get_string_attribute() now correctly converts non-string attributes to string
bmp: Correctly handle the combination of greyscale + RLE compression
dds: Corruption protection: validate resolution and guard against integer overflow
dpx: Several safety fixes for corrupt DPX files: integer overflow protection in buffer size calculations, span-based pointer safety, and use of check_open() for resolution/channel validation
heif: Fix incorrect tracking of current subimage
iinfo: Better error handling and propagation, especially from --hash; fix return code when a file could not be read
jpeg: Be more flexible with corrupt IPTC blocks; use "imageinput:strict" to control whether a bad block is skipped silently or fails the whole file
jpeg2000: Guard against integer overflow in buffer size computation
rla: Harden against corrupted files: guard against RLE buffer overruns and improve seek robustness
sgi: Better detection of corrupt RLE info that could overflow
softimage: Multiple hardening fixes against corrupted input: prevent RLE buffer overruns
targa: Protection against corrupt, mis-sized palette; fix misunderstanding of non-zero palette start index
tiff: Care with missing rowsperstrip
[8 lines not shown]
opencolorio: updated to 2.5.2
2.5.2
This is a bug-fix and security release that addresses CVE-2026-42450 and the other issues described below. It is ABI compatible with 2.5.1.
CVE-2026-42450 affects all prior OCIO 1.x and 2.x versions.
Update sysutils/intel-microcode-netbsd to 20260512
### Purpose
- Security updates for INTEL-SA-01420
- Update for functional issues. Refer to 4th Gen Intel Xeon Scalable Processors Specification Update for details.
- Update for functional issues. Refer to 5th Gen Intel Xeon Scalable Processors Specification Update for details.
- Update for functional issues. Refer to Intel Core Ultra 200 V Series Processor for details.
- Update for functional issues. Refer to Intel Core Ultra Processors (Series 2) for details.
- Update for functional issues. Refer to Intel Core Ultra Processors (Series 3) for details.
- Update for functional issues. Refer to Intel Xeon 6700 Series Processors with E-cores for details.
- Update for functional issues. Refer to Intel Xeon 6900/6700/6500 Series Processors with P-cores for details.
- Update for functional issues. Refer to Intel Xeon 6700P-B/6500P-B-Series SoC with P-Cores for details.
### New Platforms
| Processor | Stepping | F-M-S/PI | Old Ver | New Ver | Products
|:---------------|:---------|:------------|:---------|:---------|:---------
| PTL 404 | A1 | 06-cc-03/90 | | 0000011b | Intel Core Ultra Processor (Series 3)
[16 lines not shown]
newport: timing fixes, documentation fixes, XMAP code refactoring
* Update the README to describe what the newport driver really supports now.
* Refactor the XMAP9 access to mirror what I did in the newport console
driver in the kernel.
* Add some comments around the acceleration routines to do with GFIFO
tracking and potential overruns.
* Document what some of the acceleration routines are doing.
* Add an option to set the XMAP9 timing (fast, slow) in case someone's
using a 1024x768x60Hz monitor (like me).
* Always setup an RGB2 ramp regardless of 8 or 24 bit operation.
It doesn't hurt to set this up for both and it'll be useful later on
when I add/commit support for 24 bit framebuffer via hardware dithering
on XL8.
RGB2 is one of the DAC colour spaces which are used to translate the 24 bit
RGB values into what is sent to the 13W3 connector. This isn't an 8 bit
colour map for 8 bit psuedo colour, this is a gamma correction table which
the hardware can choose per-pixel by settings in the XMAP/VC2 (but is
[6 lines not shown]
(math/R) Updated 4.5.2 to 4.5.3, another update may follow,:
CHANGES IN R 4.5.3:
UTILITIES:
* tools/fetch-recommended can be used instead of
tools/rsync-recommended to fetch recommended packages into R
sources using curl on systems without rsync or behind firewalls.
PACKAGE INSTALLATION:
* C++ standard specifications (CXX_STD = in src/Makevars* and in
the SystemRequirements field of the DESCRIPTION file) are now
checked more thoroughly. Invalid values are still ignored but
now give a warning, as do contradictory specifications.
* (Preliminary) support for C++26 has been extended to Windows.
[51 lines not shown]
Restore line whose content vanished in previous
Somehow in the previous version, the content of one line was
removed, leaving just the indentation tabs... restore that line.
While here, and inspired by that line remnant, check for trailing
whitespace, and obliterate all of it that is unintentional (there
is one space in an EDIT_ME in a here-doc which is intended).
Also, in the rarely used "zones added"/"zones removed" commit message,
for the set lists, add some vertical white space before the listings
of any zones added or removed (happens so rarely, I'm not sure that
code has ever been used).
Remove unif{8,9}
They both have serious CVEs and are no longer supported.
unifi10 is supported, wip & pkgsrc has been updated to a version with the
(known) security issues fixed, and it should be supported on any platform
which can run unifi8 or 9
